ADHDTwo Words That Shrink Most Bad AI Answers
You ask AI for help because your brain is full. Then it hands you a ten-step plan with three sub-bullets each, and now your brain is more full than before you asked. That's the wall-of-text problem. The tool meant to lighten the load just added to it, using the exact format, a long structured list, that overwhelms an ADHD brain fastest.
The real problem is the default length
AI is trained to be thorough. Thorough means long. Long means more words to hold in your head, more steps to sequence, more decisions about what to do first. For most people that's helpful. For an ADHD brain trying to find one next action, a wall of text is just a new pile of crumpled papers, generated in three seconds instead of collected over a week. You didn't get an answer. You got more inbox.
This is also exactly what the research says AI plans tend to get wrong. A December 2025 review in BMC Psychology had 27 clinical experts look at AI-generated executive-function plans for ADHD. Their finding: the plans were solid in theory but weak on real-world feasibility. A ten-step plan is theoretically complete. It is almost never feasible at 9pm when you're already tired.
The worked example
Here's what that looks like in practice. Say you type into an AI: "help me get my LinkedIn profile ready." Here's a constructed example of the kind of answer that typically comes back:
"Great question. Getting your LinkedIn profile ready takes a coordinated effort across several areas, so here is a full plan to work through. Rewrite your headline to include your target job title, three relevant keywords, and a short value statement that shows what you actually do for people. Update your profile photo to a recent, well lit headshot with a plain background, ideally taken within the last two years. Add a custom banner image that reflects your industry or personal brand, sized correctly for LinkedIn's banner dimensions. Write a full About section, at least three paragraphs, covering your background, your specific skills, and a call to action for people to reach out. Go through every past job and rewrite the bullet points to focus on measurable results instead of just listing duties. Request at least five recommendations from former colleagues, managers, or clients who can speak to your specific strengths. Add relevant skills to your skills section and pin the top three so they show up first on your profile. Join five to ten industry groups related to your target field and start engaging with posts inside them. Set up a consistent posting schedule, at least twice a week, sharing insights related to your field. Turn on Open to Work settings and customize them so only recruiters can see it if you are currently employed."
That's 225 words and nine separate projects disguised as one answer. Nothing in it is wrong. All of it is too much. Now say one word back to it: "SMALLER." Here's the kind of reply you get:
"Change your headline. That's it. Do that today."
Eight words. One task. Something you can actually start in the next two minutes.
If AI does not make your next step smaller, it is not helping. It's just repackaging the overwhelm in nicer formatting.
Why SMALLER and SHORTER are the two words that matter
SMALLER cuts the plan down to one action. SHORTER cuts the words down on whatever action is left. You can use them separately or together, and you can use them at any point in a conversation, not just at the start. If AI gives you five steps, say SMALLER and it should hand you one. If that one step still comes back as a paragraph, say SHORTER and it should hand you a sentence.
This lines up with what a CHI 2026 co-design study found when it worked with 20 ADHD university students and 5 experts on what AI support actually needs to do. Three things came up: help noticing what you're actually supposed to be doing, help sticking with a task once you've started, and help with the emotional noise around both. A 225-word plan fails the first need before you even get to the second. You can't notice one task inside nine of them. SMALLER fixes that by force. It doesn't make the AI smarter. It makes the AI's output match the size of what your brain can actually hold in one hand.
None of this means long answers are always wrong. Sometimes you do need the full plan, the nine projects, the whole picture. SMALLER and SHORTER are an override, not a permanent setting. You reach for them the moment a response starts feeling like more work to read than the task itself.
Where this fits in the Yap Method
This is the same override built into TAP, the step where you pick one tiny physical action, two to ten minutes, out of everything AI sorted from your yap. SMALLER and SHORTER are how you enforce that when AI forgets and hands you a list instead of a step. No study says AI fixes ADHD, and this site won't tell you it does. What two words can do is stop a helpful tool from accidentally re-creating the exact pile it was supposed to clear.
Word count: bloated answer vs. after saying SMALLER
Same question, same AI, one word changed the output
Take the whole system with you
The free 43-page ebook: the brain science in plain words, the four steps, the master prompt, and your first seven days.
Prefer it live? Doug teaches this in libraries and meetups — book a 90-minute class →